System may have as input digital images that need to have optical character recognition (OCR) performed on them. Some systems use generic OCR for any document that comes in. That can cause numerous issues, however. For example, if the general textual structure or format of the document is not known, then the OCR may be difficult to perform and may result in misinterpretation of the data. As a more specific example, performing OCR on numerous driver's licenses from different states may result in text being determined, but it may be difficult or impossible to know what that text represents on the driver's license without further knowledge of the textual structure or format of the driver's license.
Techniques herein address these issues.
The approaches described in this section are approaches that could be pursued, but not necessarily approaches that have been previously conceived or pursued. Therefore, unless otherwise indicated, it should not be assumed that any of the approaches described in this section qualify as prior art merely by virtue of their inclusion in this section.